1. Exochorion. (Remains of the epithelium of the ovarian follicle.) 2. Chorion. (Egg-shell or cuticle secreted in the ovarian follicle.) 3. Vitelline membrane. (Primary egg-membrane. Yolk-skin or membrane.) 4. Serous or outer germ-membrane. (Serosa.) } 5. Amnion or inner germ-membrane. } Derived from the blastoderm. 6. Ectoderm. } } 7. Mesoderm. } Embryo. } 8. Endoderm. } }

In the embryo of Xiphidium and Orchelimum Wheeler has found and described with much detail a membranous structure which he calls the indusium. “The organ,” he says, “appears to have been retained by the Locustidæ, and completely lost by the embryos of other winged insects.” It arises in Xiphidium, as a simple circular thickening of the blastoderm, between and a little in front of the procephalic lobes (Figs. 511, 512, A-E), and afterwards spreads over nearly the whole surface of the egg, leaving the poles uncovered, as in Fig. 513, where it is divided into two further membranes, the inner and outer indusium, the former lying in contact with the amnion. After this the serosa “is excluded from taking any part in the development of the embryo; both its position and function are now usurped by the inner indusium.”

Hence in an egg of the Locustidæ Wheeler distinguishes, passing from within outward in a median transverse section of the egg, the following envelopes:

 

1. The chorion. 2. The blastoderm-skin-like cuticle secreted by the serosa. 3. The serosa. 4. The outer indusium. 5. A layer of dark granular secretion (probably some urate). 6. The cuticle secreted by the inner indusium. 7. The inner indusium. 8. The amnion. While envelopes 1–7 invest the whole egg; layer 8, the amnion, covers only the embryo.

Wheeler further suggests that the so-called micropyle of the Collembola (Anurida), which has been homologized with the “dorsal organ” of Crustacea, is a possible homologue of the indusium, as also the “primitive cumulus” of spiders, and the “facette” or “cervical cross” of Pentastomids described by Leuckart and also by Stiles.

The gastrula stage.—The primitive band invaginates so as to give the opportunity for the formation of the inner layer. This invagination, which at a certain stage is established along the whole length of the primitive band, forms a median furrow and may be regarded as the gastrula-invagination of insects. The lower (inner) layer thus arising afterwards spreads out under the entire primitive band (Fig. 509, B and C), the edges of which become bordered by the growing amnion-fold. (Korschelt and Heider.)

In certain forms the primitive band arises from several separate rudiments which afterwards unite. Thus in Musca and Hydrophilus the anterior and posterior ends develop first, and in Hydrophilus the procephalic lobes originate independently of the rest of the band. In the Aphides, also, according to Will, these lobes arise independently, afterwards uniting with the primitive band.

Division of the embryo or primitive band into body-segments.—Meanwhile the primitive band grows at the expense of the yolk, spreading out more and more over its surface, until in certain cases (Coleoptera, Diptera, Siphonaptera, and Trichoptera) it lies like a broad ribbon over the yolk, so that the two ends nearly meet on the dorsal side. By this time it becomes divided by transversely impressed lines into segments, which correspond to those of the larva and adult. The first of these segments is divided into two broad and flaring flaps, which are called the procephalic lobes. It becomes the antennal segment.

The mouth (stomodæum) now develops, and is situated at the anterior,^[82] and the rectum (proctodæum,) at the posterior pole, or end of the primitive band.

In Blatta, Hydrophilus, the Trichoptera, and the Lepidoptera the hindermost part of the primitive band is turned in ventrally (Figs. 534, C).

The preceding account of the relations of the primitive band to the yolk does not apply to all insects, since there are variations which appear to depend on the form of the egg, and on the amount and distribution of the yolk-masses. In certain Coleoptera, the primitive band sinks down and thus becomes immersed into the yolk. In Donacia (Kölliker and Melnikow) and Hydrophilus (Heider), and in the Chrysomelidæ and Attelabus, a weevil, as we have observed, the primitive band rests on the outside of the yolk, but in Telephorus fraxini it is immersed. In the Hemiptera it is immersed (Fig. 516), but there is much variation in this respect, the degree of immersion being most marked in the Coccidæ (Aspidiotus), and least so in Corixa. Besides the position of the primitive band, there are in Odonata and Hemiptera differences in the origin of the primitive band itself and of the embryonic membranes.

Korschelt and Heider divide the early embryo of insects into two types:

1. Into those with a superficial primitive band; viz., where there is no passage of yolk-elements into the space between the amnion and serosa. The primitive band has in such cases a relatively superficial position (Figs. 508, 509, 521, 535). Examples are certain Orthoptera (Blatta, Œcanthus, Mantis, Gryllotalpa), also certain Hemiptera (Corixa), certain Coleoptera, and the Trichoptera, Diptera, and Hymenoptera.

2. Into those with an immersed primitive band, with the space between the serosa and amnion filled with yolk (Figs. 517, 518, 534). Examples are the orthopterous Stenobothrus, Odonata, many Hemiptera (the Pediculina and Pyrrhocoris), the Coleoptera already mentioned, and Lepidoptera.

It should be observed, however, that these differences are of little phylogenetic or taxonomic value, since genera of the same order, notably the Coleoptera, differ as to the position of the primitive band, so also two orders so nearly allied as the Trichoptera and Lepidoptera.

Differences between the invaginated and overgrown primitive band.—In respect to the mode of origin of the primitive band and its relative position, there are two opposite types, though connected by transitional forms. In the one case the primitive band, i.e. its ventral portion, the “ventral plate” (Fig. 518, b, p) is pushed in or invaginated in the interior of the egg; in the other case it becomes overgrown by the folds of the amnion arising from its edges.

In insects with an overgrown primitive band, the band at the beginning is generally short and always situated on the ventral side of the egg, with the head-end looking forward, and remains in this position throughout embryonic life. There is no revolution of the embryo. The embryonal membranes arise through the formation of folds.

In insects with an invaginated primitive band, of which the Odonata afford examples, the first rudiment of the primitive band is in the form of a ventral plate of slight extent passing ventrally in the hinder half of the egg, in whose posterior section a process of invagination (Fig. 518, A, kh), soon occurs. The cavity of this invagination is the first indication of the amnion-cavity (Fig. 518, B, ah), while its wall in its thickened ventral part (K′) is concerned in the formation of the primitive band, and, in its dorsal thin part, in the formation of the amnion (B, C, am).

Revolution of the embryo where the primitive band is invaginated.—At first the head-end of the embryo is directed towards the posterior end of the egg, as in dragon-flies (Fig. 518). Also that surface of the primitive band which afterwards faces the ventral, is at first turned towards the dorsal side of the egg. In order to bring the primitive band into the later relations, there must occur the process of revolution, or turning, of the embryo. The somewhat advanced embryo of the Odonata, after the appearance of the head and thoracic appendages, undergoes a rotating motion around its transverse axis, and at the same time turns out of the amniotic cavity (Fig. 520, B). This process is so managed that near the head-region, the amnion and serosa, there closely situated to each other, are fused together, and at this place tear or burst open. Through this rent (a), in the same place in which the original invagination-opening was situated, the amniotic cavity again opens, and through the opening thus formed first the head and then the succeeding segments of the primitive band (Fig. 520, B) pass out, and remain there while the head passes on to the anterior pole of the egg on the ventral side, the embryo thus assuming a position like that of other insects. (Kowalevsky.)

In the parasitic Hemiptera (Pediculina), according to Melnikow, the opening in the membranes near the head remains permanent, and the embryo becomes everted through it, while the yolk, enclosed in the continuous membrane formed by the amnion and serous membrane, forms a yolk-sac on the dorsal surface. The same process occurs in Mallophaga, and also in Œcanthus, as described by Ayers (Fig. 521). Generally as soon as the embryo passes out of the amniotic cavity the latter soon becomes smaller and finally completely disappears.

As the embryo grows, and the sides grow up and the back closes over, the contents of the yolk-sac are soon taken up and absorbed in the intestinal cavity, which communicates with it.

In Phyllodromia, according to Wheeler, the process of revolution is “hurried through by the embryo from the beginning of the 16th to the end of the 17th day.” Several successive stages are represented in Fig. 522. In the 15th day the embryo still occupies the middle of the ventral surface of the egg. Soon the envelopes (amnion and serosa, as) rupture, an irregular slit being formed, and soon the egg and embryo are as seen in Fig. 522, B, the embryo standing out free from its envelopes on the yolk, and the edges of its dorsal growing walls (b) are distinctly marked. The tail now lies at the caudal end of the egg (Fig. 522, C). By the 17th day the walls have closed in the median dorsal line, and the embryo has grown in length to such an extent as to bring its head to the cephalic pole (Fig. 522, E).

Korschelt and Heider consider, since the primitive band of the chilopod myriopods (Geophilus) is curved in at the middle and sinks into the interior of the yolk, that in insects the invaginated primitive band is the ancestral or primitive one, the overgrown primitive band being derived from it. The overgrown primitive band by its position may also be better insured against certain mechanical attacks, perhaps also against the danger of drying up.

f. Formation of the external form of the body

Origin of the body-segments.—As we have seen, the first traces of segments appear very early, the primitive band being divided by superficial transverse furrows into segments. This segmentation into arthromeres (somites or metameres) can be observed in Hydrophilus and Chalicodoma at a time when gastrulation begins (Figs. 515, 536). The segmentation extends not only across the median portion of the primitive band, through whose invagination the inner layer (endomesoderm) results, but also to the lateral portions which become a part of the ectoderm of the primitive band. These transverse furrows correspond to thinner places in the epithelium, which in this stage forms the embryonal rudiment. It thus happens that, in the forms named, after the end of gastrulation not only the ectoderm, but also the endomesoderm, is already segmented.

So early an appearance of segmentation as that observed in Hydrophilus and Chalicodoma we must regard as a falsification of the process of development due to heterochrony. We must consider the conditions observed in other forms as the primitive ones, in which (as, for example, in Lina and in Stenobothrus, according to Graber) the gastrulation and separation of the ectoderm occurs in the still unsegmented primitive band, the division into segments occurring in later stages (Fig. 524). In these forms, then, the segmentation affects the invaginated endomesoderm, as well as the ectoderm. (Korschelt and Heider, p. 789.)

In the completely segmented primitive band may be distinguished two regions of a peculiar appearance (Figs. 515, 527), one at the anterior, and the other at the hinder end. The anterior, the primary head-section, contains the mouth-opening, and is characterized by its lateral expansions, or procephalic lobes. The other section, or posterior section, the so-called anal segment or telson, contains the anus. Between the two sections lies the segmented primary trunk-segment, which in insects consists of 17 segments. Of these the three most anterior are those destined to bear the mandibles and two pairs of maxillæ; the three following are the thoracic, which are succeeded by 10 abdominal segments, besides the 11th or telson (pygidium, or suranal plate).

It is now generally believed that there are primarily eleven abdominal segments, while Heymons has detected twelve in the embryos of Blattids and Forficula (see p. 162). In the later stages of embryonic development the number of abdominal segments is diminished, the 10th and 11th abdominal segments becoming fused. In Hydrophilus and Lina this is the case, but according to Graber, in the Lepidoptera there is a fusion of the 9th and 10th abdominal segments, the llth remaining free.

According to Wheeler, in Doryphora, and also in Chalicodoma (Carrière), between the primary head-region and the mandibular segment is interpolated a rudimentary and transitory body-segment, the premandibular segment. According to Carrière this segment corresponds to a rudimentary pair of limbs, and also to a ganglion, which participates in the formation of the œsophageal commissure (see p. 51).

The procephalic lobes.—The head-lobes, or procephalic lobes, appear at a very early period (Fig. 524, kl), before any traces of the segments of the trunk region. Ayers has shown that in Œcanthus the primitive band, in its earliest condition and before the appearance of the head-lobes, is a simple oval plate or almond-shaped thickening near the posterior end of the egg (Fig. 525, 1, 2). This plate is “soon divided into two tolerably well-marked regions by the enlargement of the head-end,” the first indication of the head-lobes (3). A depression next forms in what is to be the middle of the forehead. “It indicates the position of the future labrum, and forms the inner boundaries of the two cephalic ganglia, which are developed on either side of this depression at a much later stage.” Almost simultaneously with the appearance of this depression, two lateral folds are formed in the trunk portion of the band, which are the first indications of the maxillary and thoracic regions, the abdominal portion not yet showing traces of a division into segments (Fig. 525, 5). The thickened outer edges of the head-fold next gradually grow in towards the median line (Fig. 525, 5), and bend forward towards the region of the future mouth. The rounded angle made by the posterior end of the head-fold is the first indication of the antennæ. The embryo is now composed of four well-marked regions: cephalic, maxillary, thoracic, and abdominal. The primitive band then grows much longer, the primitive mouth and anus appear, and the appendages bud out, and eventually the embryo revolves and appears on the ventral side of the egg (Fig. 525, 6).

These primitive regions of the primitive band, before the segments are formed, are called by Graber macrosomites, and the secondary segments into which they divide (which afterwards become the body-segments), microsomites. The macrosomites are peculiar to insects, and may be an inheritance from a hypothetical ancestral form. With Korschelt and Heider, we should hardly share this view.

Our observations on locusts show clearly (1) that the procephalic lobes are the pleural portion of the first cephalic or antennal segment; (2) that the antenna is an appendage or outgrowth of the procephalic lobes; (3) that the eyes are a specialized group of epidermal cells of the upper part of the procephalic lobes, and are not homologues of the antennæ or of the appendages in general; and (4) it seems to follow from a study of the relations and mode of development of the clypeus and labrum, that they arise between the procephalic lobes, and probably represent the tergal part of the antennal segment, forming the roof of the mouth, i.e. closing in from above the pharynx.

In general the formation of the body-segments into the primitive band is in succession from before to the hinder end. This successive appearance has been observed by Graber in genera of different orders (Stenobothrus, Lina, and Hylotoma). For example, in the beetle Lina, after the appearance of the mandibular and two maxillary segments, appear the three thoracic segments, together with the two anterior abdominal segments, the other abdominal segments arising afterwards. In other cases, the formation of segments seems to be simultaneous along the entire length of the primitive band. An exception to the rule has been noticed by Heider in Hydrophilus, as in this beetle the development of the segments of the middle region appears somewhat delayed, while the fore and hind parts of the primitive band are more rapid in development. In Pieris, according to Graber, the thoracic segments are more rapidly developed than the others; soon after, the gnathic segments (mandibles and two pairs of maxillæ) appear, and finally the abdominal segments are formed.

Fore-intestine (stomodæum) and hind-intestine (proctodæum), Labrum.—The digestive canal of insects consists, as in other animals, of three portions, the fore, mid, and hind gut or intestine. The next change after the completion of the segments of the primitive band is the development of the fore and hind intestine and the appendages. The fore-intestine (stomodæum) arises as an invagination in the area of the primary head-section, and the hind-intestine (proctodæum) in the terminal section (Figs. 300 and 546).

In insects generally the formation of the fore-intestine occurs earlier than that of the hind-intestine. An exception was discovered by Graber and also by Voeltzkow in Muscidæ, where the proctodæum appears earlier.

Usually at the time of origin of the stomodæum a projection arises at the anterior edge of the primary head-region, the so-called forehead (Fig. 527, vk), which is the common rudiment of the clypeus and labrum. In many cases (certain Coleoptera and Lepidoptera) these rudiments first assume the form of paired hooks (see Figs. 83, 102, 104, 105, of Graber’s Keimstreif der Insekten, also Figs. 529 and 546), which afterwards, by fusion in the median line, become single, though notched in the middle; but in the more generalized Blatta and Mantis, as well as in bees, the rudiment is single at the outset.

The view advanced by Patten, and also by Carrière, that the labrum is a first pair of antennæ, is scarcely tenable, and we quite agree with Korschelt and Heider in regarding the clypeo-labral region as homologous with the upper lip of Crustacea, and, we may add, of Merostomes and of Trilobites.

It should be observed that in many insects, in their earlier embryonic state, directly behind the mouth arises, from paired rudiments, what seem provisional lower lip structures (not to be confounded with the 2d maxillæ of insects). This under lip structure was first discovered by Bütschli in the bee (his inner or 2d antennæ), and afterwards by Tichomiroff in Lepidoptera. Heider, in his work on Hydrophilus, describes it as the “lateral mouth-lips,” while, more recently, Nusbaum has observed it in Meloë. This under lip structure may be regarded as analogous to the paragnaths of Crustacea, although to attempt to homologize it with these seems useless. (Korschelt and Heider.)

Completion of the head.—Sufficient attention has not been paid to this subject by embryologists. The head is at first, dorsally, mostly composed of the head-lobes, or antennal segment only, and the dorsal or tergal portion of the oral appendages develop at a later period. We have observed in the embryo of dragon-flies (Æschna) that the tergites of the mandibles and first maxillæ are simultaneously fused with the head-lobes, while the much larger tergal region of the 2d maxillæ remains for some time separate from the anterior part of the head, and is continuous with the thoracic segments, and it is only just before hatching that this segment becomes fused with the rest of the head (Fig. 36). In a sense, the 2d maxillary segment when it is free from the head reminds us of the foot-jaw, or 5th segment of chilopod myriopods (see also p. 53).

g. The appendages

As we have seen, nearly or quite simultaneously all the limbs as a rule bud out from each side of the median line of the primitive band. They arise as saccular evaginations or outgrowths of the ectoderm, directed a little backwards. They are at first filled with mesoderm cells, and in the Orthoptera diverticula of the cœlom-sac are taken up into the rudimental limbs, as in Peripatus and Myriopoda. (Graber, Cholodkowsky.) As the antennæ, mouth-parts, legs, and abdominal appendages are all alike at first, their strict homology with one another is thus demonstrated. In insects never more than a single pair of limbs is known to arise from one segment.

The cephalic appendages.—The antennæ evidently arise from the hinder edge of the procephalic lobes (Fig. 527, an). As in Limulus, the first pair of appendages are at first postoral (Fig. 528, at), afterwards moving forward owing to changes in the relative proportions of the parts of the head, and they are in all respects, in their development and position in relation to the segment from which they arise, homologous with the appendages succeeding them.

The occurrence of rudiments of a pair of preantennal appendages in Chalicodoma which is claimed by Carrière, needs confirmation, as other embryologists have not observed them.

The postoral appendages of the head are the mandibles and the 1st and 2d maxillæ, besides the supposed premandibular segment already referred to on pp. 50–54, which only temporarily exists.

The trophi or oral appendages are all alike at first, but soon differ in shape, acquiring their characteristic form shortly before the embryo leaves the egg. The mandibles of Œcanthus are said by Ayers at the time of revolution of the embryo to be slightly bilobed, and in his Fig. 5, Pl. 19, they are represented as deeply trilobed, but in general they are undivided. The 1st maxillæ are at this time distinctly trilobed. The 2d maxillæ are separate, and distinctly though unequally bilobed, becoming united shortly before birth. In the embryos of dragon-flies they are at an early date very large and long, and directed backwards, and are not fused together until just before hatching, when the extraordinary mask-shaped labium is fully developed.

The distal parts of the labium, such as the ligula, palpifer, and palpus are elaborated before the mentum and submentum. Many details as to the final changes in the mouth-parts before hatching remain to be worked out.

The thoracic appendages.—The three pairs of legs arise at the same period and in the same manner in all insects; it is not until the end of embryonic life that they become jointed, and that the claws and onychia are developed. Especial attention has not yet been given to the details of the development of the parts of the last joint of the tarsus.

In many forms the antennæ are the first to appear, the mandibles, maxillæ, and legs appearing at a latter date, though simultaneously. It is thus in Stenobothrus, Hydrophilus, and Melolontha. In Lina, according to Graber, the mandibles precede the antennæ in appearance. In the Libellulidæ, according to Brandt, the legs first appear, then the jaws, and lastly the antennæ. This did not seem to be the case in the embryos of Æschna observed by us, although our observations were more superficial.

On the other hand, in those insects whose larvæ are footless, the rudiments of the legs are retarded and aborted just before hatching (fossorial Hymenoptera and Apidæ), or the rudiments of the legs are not developed at all.

The abdominal appendages.—These appear soon after the thoracic limbs, corresponding in most cases to the latter in shape and position, and their position in the embryo is a matter of the greatest interest. Von Rathke was the first embryologist to detect those of the first abdominal segment, in his examination of the development of Gryllotalpa. Long afterwards Bütschli detected them in the embryo of the honey-bee, observing a pair on each segment. Patten observed them in Trichoptera; Kowalevsky first perceived them in Lepidoptera, Tichomiroff confirming his observations. Graber, Ayers, and Wheeler have observed them in Orthoptera and Coleoptera, and the latter has detected them also in Hemiptera and Neuroptera; and while they do not arise in the embryos of Diptera and of Siphonaptera, they are to be looked for in any or all the lower or more generalized orders.

As the result of these discoveries of polypodous embryos occurring in all but the most specialized order (Diptera), it appears to be a rational deduction that the winged insects have descended from insects in which there were functional legs on each abdominal segment. Such an ancestor was the forerunner of the Thysanura, in which abdominal locomotive appendages still survive, though in a modified, more or less aborted condition. This polypodous ancestral form was apparently allied to Scolopendrella, which has a pair of functional legs on each abdominal segment.

The subject, then, of polypodous embryo insects is one of special interest, and has attracted much attention from Graber, Wheeler, Haase, and others. That these are genuine, though transitory appendages, is shown by the fact that certain pairs persist throughout adult life. The embryology of the Thysanura when worked out will throw much light on this subject, but we know that the spring (elater) of Collembola (and possibly the collophore) and the cerci of the winged insects are survivals of these limbs. That the three pairs of appendages forming the ovipositor, or sting, are most probably derived from these appendages is claimed by Wheeler (p. 167), and seems proved by the fact that Ganin and also Bugnion has detected three pairs of imaginal disks in the embryo of parasitic Hymenoptera. Hence the abdominal appendages may ultimately be found to arise in nearly all cases from imaginal disks like those giving origin to the cephalic and thoracic appendages.

As regards the Diptera, Pratt has observed that each of the three thoracic and eight abdominal segments of the embryo brachycerous Diptera (seen especially well in Melophagus) has two pairs of imaginal disks, a dorsal and a ventral pair. He thinks there is no doubt but that the ventral abdominal disks are homologous with the rudimentary appendages which appear in the embryos of all other insects, though not in the brachycerous dipters.

Appendages of the first abdominal segment (pleuropodia).—As early as 1844 Rathke observed in the embryo of the mole-cricket a pair of appendages on the 1st abdominal segment, which he described as mushroom-shaped bodies, and supposed to be embryonic gills. They are called pleuropodia by Wheeler, who, with Patten, Graber, and Nusbaum, ascribes a glandular function to them, while Wheeler suggests that they were odoriferous repugnatorial organs. In Blatta (Phyllodromia) they are of large size, in Melolontha enormous (Fig. 528, B) and filled with blood. Wheeler distinguishes as varieties, beside the mushroom-shaped appendages of Gryllotalpa and Hydrophilus, the reniform (Œcanthus), the broadly pyriform (Blatta), and the elongate pyriform (Mantis carolina). In the European Mantis they are most limb-like, with a digitiform continuation divided by a constriction into two sections. (Graber.) In Meloë they assume the shape of a stalked cup. (Nusbaum.) In the bee and in Lepidoptera the pleuropodia are not present, though the temporary appendages on the succeeding segments appear; Carrière, however, found them on the two first abdominal segments of very young larvæ of the wall-bee (Chalicodoma).

Their cellular structure is peculiar, and they are either formed by evagination or invagination, those of the latter type being subspherical and solid. Those of the former type have a cavity communicating by means of a narrow duct through the peduncle with the body-cavity (Blatta). No tracheæ, nerves, or muscles enter them, though blood-corpuscles have been seen in the cavities. “In some species the pleuropodia produce a secretion from the ends of their enlarged cells. This secretion may be a glairy albuminoid substance (Cicada, Meloë), a granular mass (Stenobothrus), a bundle of threads (Zaitha), or a thick, striated, cuticula-like mass (Acilius).” They attain their greatest size during the revolution of the embryo, and they are “mere rudiments of what were probably in remote ages much larger and more complex organs.” (Wheeler.)

Lameere has observed that in Phyllodromia the first pair of abdominal appendages, after becoming of considerable size, undergo an enlargement at their free end, become detached, and fall into the amnion.

Wheeler also calls attention to the homology of these pleuropodia with the 1st abdominal appendages of Campodea, shown by Haase to be originally glandular, but with at present a respiratory function. In the embryos of later, higher orders of insects, these appendages are in size and shape similar to those of the succeeding segments. (See also p. 164.)

Are the abdominal legs of larval Lepidoptera and phytophagous Hymenoptera true limbs?—The presence of these abdominal legs in the embryos of Sphinx (Kowalevsky), of Bombyx mori (Tichomiroff), and both Bombyx mori and Gastropacha quercifolia (except those of the first segment), as well as in Hylotoma, which has 11 pairs of such appendages, has suggested that the prop or prolegs of caterpillars and saw-fly larvæ are survivals of these outgrowths, and not secondary, adaptive structures. Opinions on this point vary. Balfour, and also, more recently, Cholodkowsky, hold that the prolegs are survivals of the embryonic appendages. Graber cautiously, after a lengthy and interesting discussion, says that the question cannot be, in the present state of our knowledge, solved. He, however, seems inclined to believe that the prolegs are not merely secondary structures, and that the rudiments of limbs may remain for a long time in a latent state before their final development. Korschelt and Heider are disposed to regard the abdominal appendages of Lepidoptera and Hymenoptera as true limbs, referring to Balfour’s statement that in the Crustacea there are different examples of the loss and later appearance of limbs, such as the loss of the mandibular palpi of the zoëa of decapods, and the loss in the zoëa of appendages in the Erichthus form of the Squilla larva corresponding to the third pair of maxillipedes and first two pairs of legs of Decapoda, and which are afterwards reproduced; similar cases occurring in the Acarina. In the wasps and bees also, as is well known, the imaginal disks of the thoracic appendages appear, the legs themselves being suppressed in the larva (the imaginal disk probably existing in an indifferent state), to reappear in the pupa and imago. It does not, however, necessarily follow that the numerous pairs of hooked ventral tubercles of certain dipterous larvæ (Ephydra) are true appendages.

It seems to us that it is a strong argument for the view that these prolegs are survivals of primitive limbs, that from similar embryonic paired outgrowths on different segments arise the spring of Podurans, the anal cerci, and three pairs of appendages forming the ovipositor, and the anal legs of the Corydalus larva, as well as those of caddis-worms; at least five abdominal segments throughout the class of insects as a whole bearing appendages in the adult.

On the other hand the view of Haase, that the prolegs of caterpillars are secondary, adaptive characters, is supported by the fact of the rapidity with which two pairs on the 3d and 4th segments nearly disappear in the larvæ of certain Noctuidæ (Catocala, etc.), a reduction evidently due to disuse.

The tracheæ.—The tracheal system arises as ectodermal invaginations on one side of the appendages, appearing soon after the latter. The earliest condition of the tracheal invagination is seen in section at Fig. 539, E, tr; as it deepens, it sends off diverticula or tracheal branches, while the narrow mouth of the invagination forms the stigma. The cup-like cavities situated serially one behind the other, and arising from the single tracheal invaginations, become at the end or bottom of the cup elongated along the length of the body and fused together at their ends; then the two longitudinal stems of the system arise, by a breaking through at the place where the original invagination had become fused, thus forming a continuous tube, the lumina opening into each other. (Bütschli.)

The cuticular tracheal intima is differentiated late in embryonic life. The entrance of the air is accomplished in part before the embryo hatches, the air being derived from the tissues and fluids of the body.

The farther development of the tracheal branches is due to the progressive formation of diverticula. The branches thus arising are intercellular formations. On the other hand, the finest twigs are intercellular structures. However, as Schaeffer states, the differences between the two modes of formation are not important.

Wheeler mentions the existence of “two pairs of very indistinct tracheal openings in the 10th and 11th somites” of the abdomen of Doryphora (Fig. 546, t₁₉, t₂₀), and Heider believes that they exist in Hydrophilus.

The tracheal invaginations as a rule begin to appear after the appendages commence to bud out. An exception is met with in the bee (Apis), where the tracheal ingrowths are seen before the rudiments of the legs. Most of the tracheal invaginations appear simultaneously. Only rarely do we see an indication of their successive development from before backwards. Thus in Hydrophilus, Graber observed that the mesothoracic stigmata appeared somewhat earlier than those of the other segments.

h. Nervous system

The rudiments of the nervous and tracheal systems essentially contribute to the building up of the relief of the primitive band of insects. The nervous system is the earliest to appear, being indicated very early, in fact before the appendages begin to grow out. The first traces of the nervous system are two ridges extending along the primitive band, the depression between them being called the primitive furrow. At an early period the segmentation is observed in the primitive ridges, while widened spaces (the rudiments of the ventral ganglia) alternate segmentally with the narrow places which are the incipient longitudinal commissures (Fig. 527, A, g).

The primitive ridges extend anteriorly into the head-lobes; this part must be regarded as the rudiment of the œsophageal commissure. The rudiments of the brain are from their first appearance directly connected with the ventral chain of ganglia.^[83]

Completion of the definite form of the body.—This is accomplished by the growth of the primitive band around the yolk, the band widening, so that its edges behind the head extend up, and finally meet on the back, forming the back or tergum of the embryo, thus enclosing the yolk (Fig. 530, F). The tergal wall of the head is due to the dorsal growth of the head-lobes, and of the clypeo-labral region. In the course of this process the anterior end of the primitive band becomes turned up dorsally, forming a dorsal curve or bend. By this bending up of the primitive band the forehead nearest the mouth forms a transverse ridge, the labrum, while the basal or earlier part of the forehead now is differentiated into the clypeus. This clypeo-labral region likewise forms the roof or palatal region of the mouth. The head-lobes cause by this dorsal growth a rotating motion which carries the rudimental antennæ back over the mouth.